> On May 12, 2015, 9:31 p.m., John Speidel wrote:
> > ambari-server/src/main/java/org/apache/ambari/server/controller/internal/BlueprintConfigurationProcessor.java,
> >  line 1396
> > <https://reviews.apache.org/r/34123/diff/1/?file=956806#file956806line1396>
> >
> >     what about double quoted style?
> 
> Robert Nettleton wrote:
>     Double-quoted flow style was not necessary for this particular patch.  
> I've added the enum for the sake of completeness, but did not want to add 
> this unless we needed it for config processing.
> 
> John Speidel wrote:
>     Seems dangerous to me to have the double quote flow available but not 
> implemented.  I can certainly see somebody utilizing it in the future because 
> it is available via the enum and not checking to see if it is actually 
> implemented.  If you don't want to implement it, it would be safer to remove 
> it from the enum.

It's important to note that this is an internal implementation class for the 
BlueprintConfigurationProcessor, and not a public-facing API.  While I don't 
agree that this is dangerous, I'll remove the enum for double quoted flow style 
in order to resolve this issue.  Once I remove the enum, I'll resubmit the 
patch.


- Robert


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/34123/#review83489
-----------------------------------------------------------


On May 12, 2015, 8:55 p.m., Robert Nettleton wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/34123/
> -----------------------------------------------------------
> 
> (Updated May 12, 2015, 8:55 p.m.)
> 
> 
> Review request for Ambari, John Speidel, Mahadev Konar, and Robert Levas.
> 
> 
> Bugs: AMBARI-11087
>     https://issues.apache.org/jira/browse/AMBARI-11087
> 
> 
> Repository: ambari
> 
> 
> Description
> -------
> 
> This patch resolves AMBARI-11087.
> 
> The BlueprintConfigurationProcessor has been modified to account for a new 
> Storm property, "nimbus.seeds", that must be set to a comma-separated list of 
> Nimbus Server hosts.  This property is used by Storm clients to discover the 
> list of HA Nimbus servers available in the cluster.  This property uses Yaml, 
> and so the default Yaml PropertyUpdater in the 
> BlueprintConfigurationProcessor has been modified to account for different 
> Yaml Flow styles.  This patch also registers this property updater to 
> "nimbus.seeds" for processing during cluster creation. 
> 
> This patch also adds unit tests to verify this change, and updates existing 
> unit tests to accomodate this change as well.
> 
> 
> Diffs
> -----
> 
>   
> ambari-server/src/main/java/org/apache/ambari/server/controller/internal/BlueprintConfigurationProcessor.java
>  7938cc1 
>   
> ambari-server/src/test/java/org/apache/ambari/server/controller/internal/BlueprintConfigurationProcessorTest.java
>  e7b0c64 
> 
> Diff: https://reviews.apache.org/r/34123/diff/
> 
> 
> Testing
> -------
> 
> 1. Manual testing done with a 3-node cluster.  Current problems with Storm 
> keep the initial Storm services from starting properly, but I was able to 
> verify that the "nimbus.seeds" configuration property is set correctly at the 
> completion of the cluster startup.  With a local patch to Storm itself, I can 
> get all the Storm services to start properly. 
> 2. Ran the ambari-server unit tests, ran into 1 un-related failure:
> "Failed tests:   
> testPopulateResourcesForRegexpMetrics(org.apache.ambari.server.controller.metrics.timeline.AMSPropertyProviderTest):
>  expected:<...rics.%25.AvailableMB[]&appId=RESOURCEMANAG...> but 
> was:<...rics.%25.AvailableMB[%2Cyarn.QueueMetrics.%25.AvailableMB._min%2Cyarn.QueueMetrics.%25.AvailableMB._max%2Cyarn.QueueMetrics.%25.AvailableMB._avg%2Cyarn.QueueMetrics.%25.AvailableMB._sum]&appId=RESOURCEMANAG…>".
>   This unit test failure will be fixed in a separate patch by the Metrics 
> team.
> 
> 
> Thanks,
> 
> Robert Nettleton
> 
>

Reply via email to